 Live from the San Jose Convention Center, extracting the signal from the noise, it's theCUBE, covering Hadoop Summit 2015. Brought to you by headline sponsor, Hortonworks, and by EMC, Pivotal, IBM, Pentaho, Teradata, Syncsort, and by Atunity. Now your hosts, John Furrier and George Gilbert. Okay, welcome back everyone. We are here live in Silicon Valley for Hadoop Summit 2015. This is theCUBE, SiliconANGLE's flagship program, where we go out to the events and extract the signals from the noise. I'm John Furrier, the founder of SiliconANGLE. Join my co-host. This week our new big data is George Gilbert at Wikibon.com. Our next guest is David Richards, founder and CEO of WAN Disco, CUBE alumni, been on theCUBE many times. Great to see you again. Good to see you guys again. Always a pleasure. Public company, so we're going to get all the secrets out there. I know you like to reveal all your secrets to your investment, I'm only kidding. Yeah, the CFO is not here, I'm off the shuttle. So earnings, give us some guidance. So you guys are great, been big supporters of theCUBE. Certainly WAN Disco, we had a great relationship with you guys, know all your folks technically and through the business. But the business is actually spinning in your direction right now. The market forces is everything. You guys made the good decision to drop the distro, all the stuff, the tooling you're working on and doing the replication, multiple data centers, enterprise grade. You guys have been on this path and haven't wavered. Now it's trending today. So tell us what's going on. So that's a really good summary actually of where we are. So the heritage of WAN Disco of course is in replicating data, replicating stuff, replicating whole data centers. And we placed a big bet two years ago when we could see Hadoop coming over the horizon and disrupting, fundamentally disrupting the storage space. And I know George, you do a great presentation about that storage gap between budgets and cost and storage growth. And we could see that happening. What we could also see happening was enterprises adopting this stuff. And if enterprises adopt this stuff, there are a couple of things that they fundamentally need. They need some way to manage it. The guys that stare at screens all day, they need that single pane of glass as they call it. They need security products, because banks, especially right now, security is probably the number one concern that they've got. And it also needs to be super reliable and resilient for enterprise use. And that was the bet that we placed with our unique technology. So the first iteration of the product actually was called non-stop Hadoop. And in order to do that, we had a modified name node. So we modified the name node. So we had to modify the underlying, we had to work very closely with Hortonworks, Ladiere and all those guys. In order to modify their distributions. The new product that we've just launched, which is going just thin, and I'm not giving any guidance here, but the new product called Wandisco Fusion, is just knocking the ball out of the park. Because it doesn't rely on any changes to the underlying infrastructure. It provides tons and tons and tons of benefits. Everything from replicating data, super reliability. Not like an Oracle Fusion project. That took eight years there to write. And still, yeah, still. Why is it successful? Because we're hearing that it's not a one-trick pony world anymore. I'm going to have Hadoop here. I'm going to have a lot of stuff going on in the enterprise. And the enterprise guys don't buy that way. They don't say, give me some MapReduce, give me some Hadoop. They talk in different language. Is that the main reason that you guys are winning with Fusion? Can you just elaborate on that? So for a couple of reasons. One is the resiliency, reliability of Hadoop in general. The other thing is this is not going to be a one-size-fits-all marketplace. When you say one-size-fits-all, you mean the one ring to rule them all in terms of one cluster. It's not one cluster. Correct. And there are going to be different storage arrays. So it might be one division of a large bank that we did an implementation with recently was using Hortonworks. Another division was using Cloud Era. Another division was using for dense archival EMC Isilon. And then they want to burst out to the cloud. Now that in itself creates a problem. And that's the problem that Fusion solves. So we can take all of that. Yeah, you integrate and unify. Unify, integrate, any type of HCFS compliant storage. So the problem space is the customers sitting there and they built out the Isilon for the NFS reason. They have these solutions there for business outcome reasons that they paid some cash for. Now their problem is we've been successful with them. We got to go to the next level. So their choices are throw it all away or replatform everything or rip and replace or put you guys in and keep everything. And did I get that right? Yeah, all the other thing that we're seeing and I could talk forever about all the advantages of this but one of the other main drivers for this is migration. Now we've only just, it's like 10 minutes ago we stepped into the Hadoop marketplace but believe it or not a vendor that's got and I won't be disrespectful to any of the Hadoop distribution vendors that has Hadoop distribution A wants for various reasons to conglomerate that onto Hadoop distribution B. When you've got petabytes and petabytes of data you cannot take a production system that using Hadoop and migrate to something else because you're going to have an outage in downtime and you're going to lose data. Well, guess what? When this goes active, active replication between different distributions you can do precisely that. You guys really hit a good bet there. Yeah. I don't want to geek out but it's also you know there's this Hadoop 2.0 was supposed to be really about yarn and so there's this notion that oh I can run all my workloads on now one cluster but that's not entirely true is it? No, I think it's not and we can geek out George I'm perfectly okay with that. So I think what Hadoop 2.0 is driving towards Hadoop 1.0 was really, it was the legacy of big table from Google which was really designed as a batch processing system to process the internet and I know you guys have gone through this story a million times. Hadoop 2.0 really was the move from it being a batch processing system to a compute platform and we're seeing yarn, we're seeing Spark, we're seeing all of those other compute tools now sitting on top of Hadoop and Hadoop really morphing from being batch processing to real time compute processing. So the other thing we're hearing too is it's not just about Hadoop, there's other things going on in the data center so I might want to run a different platform with Spark for instance. Spark now is in Hadoop. So this is kind of some new things. What's your take on that trend and how does Hortonworks fit into this? I call post Hadoop growth mode not Hadoop shrinking. Certainly Hadoop's growing, we see that. But there's other stuff outside of the size of Hadoop that might sit at the analytics layer. It's very easy for us all to get massively carried away right with the Hadoop marketplace. But when you look at the actual, in relative terms the revenues associated with this marketplace these guys are circa a hundred million dollars. If you go look at somewhere like Cisco where the UCS division in three years created a three billion dollar company. IBM. IBM, all these guys, they're multi-billion dollar companies. We sometimes, the bright shining stars and the bright shining lights we sometimes just get a little bit carried away. But as you just said, rightly. In heat inside IBM runs the big analytics group over there. I mean Hortonworks annual revenues is like a rounding number on her spreadsheet. That's right. I mean from an IBM scale standpoint but they have different customer base. That's the reality of the world though. The customers themselves, they're going to buy Hadoop. They're not thinking 24 seven about Hadoop. Well I mean look, what's Hadoop actually doing to the marketplace? It's saying I'm going to commoditize storage. I'm going to solve this problem where you're going to need to store petabytes and petabytes of data that you couldn't previously store and you simply can't use proprietary expensive storage arrays to do that. You wouldn't want to do that. But that doesn't mean to say that those proprietary expensive storage arrays are going to disappear into the sunset. I mean look, one of our biggest lead generating partner for us, believe it or not, it's Orital. It is. It's Orital's BDA. The big natural appliance from Orital. Oh right. Those guys are killing it. Now they're arguably one of the biggest Hadoop vendors in the marketplace that we never talk about. So it's not just about those bright shining stars. It's a packaging issue. It's a, you know. So let's talk about that. That's packaging. This is exactly what we were talking about with Murray, Adrian Gardner. We're high-fiving each other in the industry saying we're crossing the chasm. But the reality is the customer's like, no, you hit the ball in the water now. Drop it in the drop area. You're still in play, but you're not out of business. But you still got to, there's a lot more work to do than nail the customer experience. And we've kind of seen this movie before where we went from mainframes to three tier client server. And you know, some of the biggest vendors in that, in the early part of the three tier client server marketplace were guys that could make mainframes work in, you know, screen emulation technology could work in a three tier client server marketplace. Yes, it does create new vendors like it did back then, SAP and so on. And yes, big data will create new vendors like Hortonworks and Cloudera that will eventually become, I'm sure, multi-billion dollar, very successful companies. But in this early part of the marketplace, we see a great deal of demand, particularly from financial service institutions, banks, who really don't want to run their own data centers for a start. I mean, they kind of have to at the moment because of all the compliance and security issues. But number two is they really don't want to implement software. They want to buy an appliance from somebody. So actually, the appliance vendors, I think, are a little bit behind the curve here. They should be doing a lot better than they currently are. And it makes sense too, because one of the things we talk to app developers are saying is, hey, I'm writing an application on big data, whether it's a Dupa or whatnot. And as I become successful, they have domain expertise, they're making money, and as they go down to the customer base, they have to write more software. Because like, oh, we have a systems management product, you've got all these legacy software connectors, there's a problem. Yeah. The guys in entrepreneurs want to write more software. So an appliance or a platform that can do that work is what we're seeing happening. Look, I always imagine this kind of like ship that's sailing across the ocean. If the waters are choppy and things are changing underneath all the time, big financial services companies hate that. They don't want to do that. It's very, very expensive. And it's also, it creates risk in their organizations. They can buy an appliance that's super reliable that they can just turn it on. No premium, no premium. They will pay it. They pay for sanity. Yeah. So that they, like they would want exadata if it weren't so sort of locked in and expensive. I think so, yeah. And so if Oracle offers, you know, exadata in the cloud and the pricing were reasonable. Well, Oracle kind of are with that big debtor appliance, the BDA, right? But that's tied to exadata, I think, no. No, not necessarily. We've seen it being used independently. Oh. Let's talk about the next step for you guys. I want to ask some questions. You brought this earlier and I like this. You made the good bets. It's a good team, strong management team. What's the next bet? I mean, let's get the... I've got goosebumps going down by spine at the moment because the next step for us is, okay, at the moment we're reliant on HCFS, the Heddyp Compatible File System API, which is pretty big marketplace. It's EMC Isilon, it's Teradata, it's SaaS, it's Hortonworks, CloudEra, Pivotal and all those guys. The next step for us is any storage system. The Fusion platform, as it is today, has a very thin veneer that allows us to integrate with HCFS compliant file systems. Imagine if we could do this with NFS. Imagine if we could do this with SQL databases. Imagine if we could, and I was with a very large UK government organization two weeks ago. They have active, active, Wanscope replication problems today that they've had for 20 years with all those systems. Could you have heterogeneous databases? Absolutely. Wow. You bring up an interesting point. Again, we love to riff on the future scenarios and you guys have great vision now. It's like, I see that. But it brings up an issue. Internet of Things is another market that is new stuff happening that they couldn't do before with Big Data. So Big Data creates some problems with them successes. So success can create problems. Yeah. You grew too fast. Well, so the great thing, you know, Incarnation One of WANDISCO's Big Data Play was to be very, very close to the Hadoop distributions. Incarnation Two was to take a step back and take a proxy approach without any changes, without any reliance on any changes required to those back end systems. And that's the key for us because it can take us, not just for Hadoop, but beyond Hadoop and into any kind of file system. So what other bet specifically besides the, that you see around the horizon? Or I shouldn't say bet. I won't say hard bet, like, but just your eye on the path, you see, is it a straight and narrow for WANDISCO or is there some courage you got to manage and negotiate and maybe slow down and take quietly? So you've got your game on our CTO a bit later and I'm sure your game will talk about this. We already have demand. Customers saying, yeah, we're going to use, for some of our workloads, for a lot of our workloads moving forward, we're going to use Hadoop. But we are going to keep Icelon, we are going to keep all these other things that we've already got in our data centers. But we would love it if you could also do this with NFS, if you could also do this with, you know, side-based. So you guys want to be the multi-vendor platform for connecting data and making it smart. We want to be the de facto replication layer across all storage, not just Hadoop storage across all storage. Now that takes our addressable market to, you know. That's your point about IBM and Oracle the size of these other whales. But that's the reality of customers. I've got to ask you a question. You do a lot of customer visits. I want to ask you some pointed questions on customer linguistics. What language are they talking about? Because, you know, we talk speeds and feeds, we're talking about yarn, Imbari, MapReduce, you know this, I still want to invest, forget products and technologies for a minute. How does that translate to the customer language? What are they speaking? You know, they're speaking, you know. I'll tell you a story and I wish I could. Outcomes and business transformation, that sounds cliched, but I mean, be specific, what are you hearing? Custom conversations. I wish I could tell you who this was, but I'll try not to. So four years ago, the co-founder of one of the big Hadoop distribution companies went into a Wall Street bank and told them how they were going to kill Oracle and kill Teradata and destroy all these companies. Doesn't take much to guess which. And it absolutely scared the pants. They wanted the guy out the door, that they absolutely scared. The customer. Yeah. The customer wanted the vendor out the door because it was scaring them half to death because they are not, repeat not, going to suddenly say, you know what, this new thing, I'm just going to rip it up and throw it out the window. And then I'm going to take a look. And this is the start of end of doing a deal with Oracle. It is. Yeah. Now, now it's. That's clod era. Yeah. So I'm not going to say who it is. No, no, I'm not saying. But that is, you know, that's where we, we all kind of got it wrong two years ago. The icing on the cake, the hyper growth is in those new storage arrays that are probably going to be Hadoop, but they are not going to necessarily rip and replace existing storage cases for Hadoop. No brainer. Correct. Where Adrian laid it out, almost half the enterprise survey said they will be planning on doing Hadoop. That's not, that's huge. I mean, half the, half. But how big is that? We don't know. Exactly. And if you, if you look at the, you know, the revenues in the market today and you compare that to the revenues of existing storage vendors in the market today, there is a big disparity which suggests that it's Hadoop and something else, which is why we design Fusion. Awesome. So what's next? What's going on with you? Tell us what's happening with you personally, business, what's traveling a lot? Share some stories. Tell some David Richard stories. I, I think, I think the British Airways cabin crew kind of know exactly what I like to do. Seems to bring me a drink as I, as I jump onto a plane. No, I look, I'm in New York next week, spend a lot of time, obviously, we're a public company with, with fund managers. And it's interesting in, you know, fund managers are trying to get their arms around this. They all want a exposure to Hadoop, you know, and particularly, you know, looking at the, the highly successful IPO that Hortonworks did, you know, you, I'm not sure we can call them unicorns when there's 136 of them now. You know, they seem to be everywhere. Well, oh, you mean, and the other 75 or deco coins. Yeah, yeah, exactly. You know, the news today, Spotify raised huge money. Their valuation's at 8.5 billion. And Uber's speculating at 50 billion. Well, no one knows the business models, the ultimate business models, that's what's crazy. I was with the presidency of, of a fairly big hedge fund slash bank two weeks ago. And we were discussing a couple of things. The London, the London property market, which is a huge bubble, which, which he likened to- The real estate bubble? The real estate bubble in London, which he likened to going to Hong Kong and seeing all those jewelry stores and why the 70 jewelers, why did they used to be 70 jewelers stores in Hong Kong? It's because people used to go into money there, the Chinese, et cetera. And the London's kind of allegedly a little bit like that, where you get people from the Far East, et cetera. It's paying crazy amounts of money, you know, like $150 million for an apartment and things like that, right? Crazy, crazy. But we also have a bubble in private equity. We have the original too much money chasing too few goods thing going on. Interest rates being where they are, this will not last. There's a bubble in venture too. The only bubble is the, not bubble is the series A market. Yeah, if you've got money, where'd you put it? Because you can't put it in a bank. I mean, the banks currently can't hold cash. It costs them money to hold cash. So the feds are talking about, you know, raising those interest rates. We will see if that precipitates a little bit of a burst of that bubble, because some of those valuations are crazy. You know, we've seen this before, George, you when you were a financial analyst, you'll have seen this before. I remember it. You'll remember it well. I know. That's concerning. So I got to ask you about this ecosystem. Let's look back where we were, founding of Hadoop, and then the stuff you laid out is a whole different narrative. We're hearing enterprise, operational integration, operationalizing Hadoop with other stuff is going to be a challenge. A lot more software grid. You got cloud on the horizon booming. The engine of innovation with Amazon, Azure, Google Cloud, that's to name a few, and then you got VMware and the enterprise, everyone else, you can see. Cloud is the engine that will just, we believe, turbo charge analytics and it's going to change the data center, which you'll be impacted by. So with all that being said, where has Hadoop come from and where's it going, this industry? How big will it be? What's going to happen to all these people behind us? Will it be more, will it have to die? What's your take? So that's really good analysis actually. So the cloud will turbo charge, as you put it, this marketplace, it will, it is the big shift. And we all know that cloud, I mean, we all pointed, look at salesforce.com and their valuation and so on. But that will be the switch that particularly, when I see every single CIO of every single bank doesn't want to run a data center. Is it a skills thing and the complexity of Hadoop? It's everything. It's labor intensive. They don't want thousands of people working in that, in their IT departments. It's a skills issue. It's a competitive marketplace. It's all of those things. They would much, much, much. I was talking to yesterday, somebody from one of the world's largest banks who was saying, we're cutting costs and we're going to outsource this. We're going to outsource most of our IT to third parties. I think they'd like to push the whole data center out to a third party and have it completely managed. And keep the app leveled. Yes, keep their differentiation because, you know, it's the core context thing and they see data centers as contextual. And that's where cloud really begins to happen. All right, David Richards, thanks so much for coming on theCUBE. We've got to wrap, give you the final word for the folks watching who aren't here. What's happening here at the event? What is the main theme at Hadoop Summit 2015? What should be the takeaway for them? What's the vibe? What's the sizzle? What's the stake? What's happening? So we're quite fussy about the way that we take leads on our booth. We've got quite a big booth here. And last year, I think we, day one, we got a hundred. This year we got 400. Now we're a pretty good gauge for where the Hadoop market is. If people really care about their data, they're going to come and talk to us at some point. So that uptick in people now really thinking about taking Hadoop and using it for not just production, not just for Twitter feeds and weather forecasts and other things, but now for mission-critical real-time analysis of real data, doing real things in their organizations. I think my takeaway would be Hadoop's getting real. It's real. Hadoop's getting real. David Richards, CEO of WAN Disco. This is theCUBE. We're real all the time, always live. Wall-to-wall interviews every day. Here, three days at Hadoop Summit 2015. We'll be right back with more after this short break.